Targeted Petroleomics: Analytical Investigation of ... - ACS Publications

May 21, 2014 - Department of Earth, Ocean and Atmospheric Science, Florida State ... Schools of Biology and Earth & Atmospheric Science, Georgia Insti...
41 downloads 4 Views 3MB Size
Download PDF
Article pubs.acs.org/EF

Targeted Petroleomics: Analytical Investigation of Macondo Well Oil Oxidation Products from Pensacola Beach Brian M. Ruddy,† Markus Huettel,‡ Joel E. Kostka,§ Vladislav V. Lobodin,∥ Benjamin J. Bythell,∥ Amy M. McKenna,∥ Christoph Aeppli,⊥ Christopher M. Reddy,⊥ Robert K. Nelson,⊥ Alan G. Marshall,†,∥ and Ryan P. Rodgers*,†,∥ †

Department of Chemistry and Biochemistry, Florida State University, 95 Chieftain Way, Tallahassee, Florida 32306, United States Department of Earth, Ocean and Atmospheric Science, Florida State University, 117 N. Woodward Avenue, Tallahassee, Florida 32306-4320, United States § Schools of Biology and Earth & Atmospheric Science, Georgia Institute of Technology, 310 Ferst Drive, Atlanta, Georgia 30332-0230, United States ∥ National High Magnetic Field Laboratory, Florida State University, 1800 East Paul Dirac Drive, Tallahassee, Florida 32310-4005, United States ⊥ Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, 266 Woods Hole Road, Woods Hole, Massachusetts 02543-1050, United States ‡

S Supporting Information *

ABSTRACT: Of the estimated 5 million barrels of crude oil released into the Gulf of Mexico from the Deepwater Horizon oil spill, a fraction washed ashore onto sandy beaches from Louisiana to the Florida panhandle. Here, we compare the detailed molecular analysis of hydrocarbons in oiled sands from Pensacola Beach to the Macondo wellhead oil (MWO) by electrospray (ESI) and atmospheric pressure photoionization (APPI) Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) to identify major environmental transformation products of polar, high molecular weight (C>25) “heavy ends” (high-boiling species) inaccessible by gas chromatography. The petrogenic material isolated from the Pensacola Beach sand displays greater than 2-fold higher molecular complexity than the MWO constituents, most notably in oxygenated species absent in the parent MWO. Surprisingly, the diverse oxygenated hydrocarbons in the Pensacola Beach sediment extracts were dominant in all ionization modes investigated, (±) ESI and (±) APPI. Thus, the molecular-level information highlighted oxygenated species for subsequent “targeted” analyses. First, time-of-flight mass spectrometry analysis of model compounds attributes the unusually large oxygen signal magnitude from positive electrospray to ketone transformation products (O1−O8 classes). Next, negative electrospray mass spectrometry reveals carboxylic acid transformation products. Two-dimensional gas chromatography with mass spectrometry analysis of anion-exchange chromatographic fractions unequivocally verifies the presence of abundant alkyl ketone fragments in sand extracts, and FT-ICR MS analysis reveals the distribution of high-boiling ketone, carboxylic, and higher numbered (3+) oxygen-containing transformation products too polar to be analyzed by gas chromatography. The results expand compositional coverage of oxygen-containing functionalities beyond the classic naphthenic acid type species to complex/mixed ketone, hydroxyl, and carboxylic acid classes of molecules that have been recently identified in produced water, emulsions, and petroleum production deposits.

■

INTRODUCTION In the months after the Deepwater Horizon explosion in April 2010, and with oil release estimates approaching 5 million barrels (170 ± 40 million gallons),1 the pristine beaches of the Northern Gulf of Mexico became a subject of widespread concern. Surface slicks would eventually oil an estimated >1000 km of beaches.2 A large body of literature addresses weathered oil on shorelines,3 and most recently, Aeppli et al. noted extensive oxidative modification of the Macondo wellhead oil (MWO), increasing with time in surface slicks, rock scrapings, and oiled sand patties.4 However, the identity of the oxidative weathering transformations was not explored in many of these low-resolution spectroscopy and GC studies due to their inherent compositional complexity and high boiling point. Thus, new analytical methods are required to address the highly oxidized petroleum transformation products. GC remains an © 2014 American Chemical Society

excellent choice for petroleum applications, namely, for light distillate analysis, evaporative weathering, and biomarker ratio measurement for source identification. However, GC analysis is not optimal for persistent, petroleum transformation products for two reasons: (1) it is limited to constituents below ∼C45, many of which are completely lost as a result of evaporation (unweathered crude oil hydrocarbons can extend well beyond C100), and (2) transformation products produced by abiotic and biotic environmental processes tend to be oxidized, and thus high-boiling, and adhere to the column and result in reduced (or no) separation efficiency. Therefore, although GC-based methods are extremely useful for the evaporative component of Received: February 21, 2014 Revised: April 28, 2014 Published: May 21, 2014 4043

dx.doi.org/10.1021/ef500427n | Energy Fuels 2014, 28, 4043−4050

Energy & Fuels

Article

Figure 1. Compositional analysis of low- (left) and high-boiling (right) petrogenic species: GC×GC/FID chromatograms for (upper left) Macondo well oil (MWO) and (lower left) June 30, 2010 Pensacola Beach sediment extract (5.5−7.0 cm depth). Positive ESI 9.4 T FT-ICR broad-band mass spectra for Macondo well oil (upper right) and June 30, 2010 Pensacola Beach sediment extract (5.5−7.0 cm depth) (lower right).

properties, due to analytical limitations associated with nonvolatile, oxygenated residues. The Aeppli study concluded that the molecular composition of weathered oil could be obtained by ultrahigh-resolution mass spectrometry. Here, we apply FT-ICR MS methodology to identify oil degradation products and the chemical functionality of oxygen-containing transformation products from petrogenic contaminants that resulted from the Deepwater Horizon oil spill. The methods and results facilitate expansion of the current application to naturally occurring, highly polar petroleum species in emulsions, production deposits, and produced waters.

petroleum weathering, they leave the environmentally persistent, transformation products largely undetected. FT-ICR MS5 is particularly suited to investigate the compositional complexity of transformation products from an oil spill because of its ultrahigh resolving power6,7 and mass accuracy8 and the ability to analyze nonvolatile and/or highly polar acidic9,10 and basic11 species. The technique also provides for high-resolution isolation,12 the flexibility to perform targeted dissociation experiments,13 and gas-phase reaction chemistry,14 tools that enable access to structural information for complex mixtures. During the past decade, FT-ICR MS has spawned the term “petroleomics” (namely, predicting the properties and behavior of petroleum and its products from comprehensive organic molecular composition)15 and enabled the petroleum industry to compositionally characterize the “heaviest” fraction of crude oils: the high molecular weight, highly polar fraction posited to be responsible for aggregation, upstream production, and refinery problems.7,16 Recently, FTICR MS uniquely identified more than 30 000 acidic, basic, and nonpolar unique neutral elemental compositions from the Macondo (Deepwater Horizon) well crude oil.17 Early environmental petroleum applications of FT-ICR mass spectrometry probed petroleum weathering in sediment and noted oxygen transformation and persistent detection of higher molecular weight species.18,19 Those results are consistent with the recent work by Aeppli et al. for the Deepwater Horizon oil spill, concluding that oxygenated hydrocarbons are oil transformation products, as confirmed by radiocarbon analysis, with a bulk oxygen content > 10%, and contain carboxylic acids and alcohols.4 However, those conclusions were based on bulk

■

EXPERIMENTAL METHODS

Samples and Preparation. The sample was collected from a site (30°19′32.08″N, 87°10′30.55″W) near Pensacola Beach, Florida, on Santa Rosa Island, a barrier island in the Northern Gulf of Mexico on June 30th, 2010 at a depth of 5.5−7.0 cm (Figure S1, Supporting Information). The site is located on a short transect in the supralittoral to the sublittoral zone. A representative environmental blank (control) was acquired from the corresponding zone of a beach on Saint George Island, FL, a barrier island ∼140 miles east by southeast of Pensacola Beach, FL, with no reported MWO contamination. The analytical blank (control) was a high purity sand sample from Ottawa, Illinois. Sand samples were stored frozen at −23 °C. Approximately 40 g of sand was mixed with an equal weight of sodium sulfate drying agent, loaded into a 30 × 100 mm cellulose thimble, and Soxhlet extracted with dichloromethane for 4 h according to American Society for Testing and Materials (ASTM) D 5369-93(2008)e1 (Standard Practice for Extraction of Solid Waste Samples for Chemical Analysis using Soxhlet Extraction). The solvent extract was then evaporated with a gentle stream of nitrogen gas to dryness, and the dry extract was

4044

dx.doi.org/10.1021/ef500427n | Energy Fuels 2014, 28, 4043−4050

Energy & Fuels

Article

weighed. Experimental details for the analytical methods employed can be found in the Supporting Information.

■

RESULTS AND DISCUSSION Complementarity of GC×GC and FT-ICR MS Analyses. An oil’s chemical composition changes after release into the marine environment.20 To provide molecular-level insight into those changes for the MWO, we employ GC×GC-FID and FTICR MS to identify the volatile and nonvolatile components of MWO weathering. For the volatile species, Figure 1 shows the GC×GC-FID traces for the MWO (top left) and the Pensacola Beach sediment extract, PB 5.5−7.0 cm, which was the most heavily oiled (bottom left). Comparison of the two chromatograms reveals that (a) the MWO (a light crude oil) has completely lost the low-boiling species (highlighted by the white oval); (b) one- and two-ring aromatics are also heavily depleted (red dashed oval); and (c) the high molecular weight alkanes and cycloalkanes remain in the Pensacola beach sediment, along with the sterane and hopane biomarkers (white dashed oval). The exceptional resolution afforded by GC×GC separation allows direct comparison of the hopanoid biomarker regions without potential interference from species co-eluting in the first dimension. The highly similar hopanoid biomarker relative abundances (and ratios) for both samples (Figure S2, Supporting Information) confirm the source of the beach contamination as the MWO. The same technique has previously been used to fingerprint the Macondo well as a source of weathered oil.4,21 The compositional analysis of high-boiling species is addressed in Figure 1, showing the broad-band (full molecular weight range) positive ESI 9.4 T FT-ICR mass spectrum for the MWO (top right) and the most heavily oiled sand sample extract, collected on June 30, 2010 (PB 5.5−7.0 cm, bottom right). In contrast to GC×GC, which accesses up to ∼C35, the broad-band FT-ICR mass spectra for the MWO and the sand extract span the identical molecular weight range (200 < m/z < 1200), with no depletion of low molecular weight species observed in weathered oil (due to their high polarity and thus high boiling point). Those components exhibit high heteroatom (N, O, and S) content and are thus not amenable to gas chromatographic analysis. Here, the number of peaks of signal magnitude ≥ 6 standard deviations (σ) above the baseline noise from m/z 210−1190 for the sand extract (32 232 ± 488) is more than 2.3-fold higher than that for the MWO (13 700 ± 80), based on triplicate experiments. In summary, weathering processes in the Gulf of Mexico led to a more than doubling of the isobaric compositional complexity of the crude oil. Prior one-dimensional gas chromatographic analysis of the highly weathered beach sediment extracts revealed a depletion in polycyclic aromatic hydrocarbons and a 1−2 order of magnitude decrease in the ratio of lighter (C6−C16) to heavier (C16−C35) alkanes.22 That behavior is similar to that for oilsoaked sand patties that were washed ashore on the coast of the Gulf of Mexico, where evaporation and dilution, as well as biodegradation and photooxidation, were identified as weathering processes.4,23 Increase in Complexity from the Macondo Well to the Beach. The higher compositional complexity can be more effectively visualized for high-resolution FT-ICR data through mass scale-expanded segments (Figure 2) that span a fraction of a single nominal mass (0.2 Da at m/z 500). For positive ESI, 10 peaks are detected for the MWO (top) and 32 for the contaminated beach sediment extract (bottom), whereas no

Figure 2. Positive ESI 9.4 T FT-ICR mass scale-expanded segment at nominal m/z 500 for (top) Macondo well oil (MWO) and (bottom) June 30, 2010 Pensacola Beach contaminant (5.5−7.0 cm depth). Contaminant species not present in MWO are shown in red.

peaks ≥ 6σ are observed from either the environmental or the analytical blank across the same mass segment (data not shown). Peaks detected from the Pensacola Beach sand extracts but not detected from the MWO are highlighted in red and expose the primary source of the increased complexity as oxidation. We assigned molecular formulas to peaks of magnitude ≥ 6σ above baseline noise and grouped the compounds according to heteroatom class. Positive ESI FTICR MS heteroatom compositions (Figure 3, top, not including carbon and hydrogen) show that the MWO heteroatomcontaining species (green) are dominated by pyridinic (6membered ring nitrogen) N1 and N1O1 heteroatom classes as well as the O1S1 (sulfoxide) class. For the contaminated beach sediment extract (red), more highly oxygenated N1Ox species are detected, in addition to the N1 and N1O1 species present in the MWO. Note that the contaminated sand also contains Ox components (x = 1−3) of even mass, seen in Figure 2 as 13C isotopologues.24 Most importantly, Ox species are not typically observed by positive ESI of raw crude oil11,18 but are sometimes observed in produced water and produced water deposits,25 because Ox functional groups in crude oil (present as ethers and esters) are difficult to protonate. However, here, we unambiguously verified the assignment of Ox species from 18 O isotopic fine structure based on stored waveform inverse Fourier transform (SWIFT) selection to extend FT-ICR dynamic range,24,26 as discussed in more detail elsewhere.27 Overall, high resolution and mass accuracy facilitate the unique assignment of elemental compositions for complex weathered petroleum samples and demonstrate the power of FT-ICR MS. However, the ultimate power of the technique is that it provides detailed compositional information for nonvolatile species not detectable by conventional GC or GC×GC MS for sediment samples collected from the same sites at Pensacola Beach.22 Most of the oil spill transformation products detected in the Pensacola Beach sand extracts are oxygenated aromatic hydrocarbons that span C20−C85. The FT-ICR-MS heteroatom class (>0.5% relative abundance) results for the Macondo crude oil (green) and Pensacola beach sand extract (red) from June 30, 2010 for ions of both polarity (±) are shown for ESI (Figure 3) and APPI (Figure 4) ionization modes. Also shown 4045

dx.doi.org/10.1021/ef500427n | Energy Fuels 2014, 28, 4043−4050

Energy & Fuels

Article

Figure 3. Macondo well oil (green) and June 30, 2010 Pensacola Beach sediment extract (5.5−7.0 cm) (red) heteroatom class distributions for (±) ESI. Pensacola Beach sediment extract oxygen class isoabundance-contoured double bond equivalents (DBE = number of rings plus double bonds to carbon) vs carbon number plots are shown below each heteroatom class distribution for both (+) and (−) ESI.

observed as negative ions are likely carboxylic acids (note the sudden onset at O2 for ions generated by (−) ESI). The sudden onset at O2 for (−) ESI and (−) APPI has previously been reported in the analysis of petroleum.10,29 Broadly distributed highly oxygenated species (with an onset at O1) from (+) ESI have been previously reported only in produced water and produced water deposits (water produced during petroleum refinery processes).25 The (±) APPI FT-ICR MS results (Figure 4A,B) largely corroborate the (±) ESI results but provide added compositional coverage of highly oxygenated, acidic species (Figure 4B, bottom) from O5 to O8. Combined, the results identify thousands of alkane, cycloalkane, and aromatic oxygenated species ((+) ESI and (+) APPI) and thousands of additional acidic alkane, cycloalkane, and aromatic oxygenated transformed species ((−) ESI and (−) APPI). Identity of Oxygenated Species Detected by (+) ESI FT-ICR MS. Because oxygen appears to account for the observed differences in the FT-ICR MS spectra, and because ketones, esters, and alcohols cannot be distinguished by mass alone, we compared various oxygenated model compounds to determine their ionization efficiencies relative to pyridinic species (native to petroleum) and to determine the chemical nature of the oxygenated species. An example of the higher signal abundance of ketones relative to alcohol and pyridinic nitrogen functional groups (at equal molar concentrations) is shown in a positive ion TOF mass spectrum (Figure S3, Supporting Information). The 10-fold higher relative abundance for a ketone (here, 4,4′-dimethylbenzophenone) relative to a phenol (1-naphthol) at equimolar concentration indicates

in Figures 3 and 4, below each heteroatom class graph are the Ox class isoabundance-contoured double bond equivalents (DBEs) vs carbon number plots for the Pensacola Beach sand extract. The additional Ox classes for the Pensacola Beach extract reported in Figure 2 extend throughout the full-range mass spectra (up to molecular weight 1200, corresponding to ∼85 carbons) from which Figures 3 and 4 were constructed. Apparently, extensive oxidative weathering converted hydrocarbons to Ox species hypothesized earlier, based on bulk measurements by Aeppli et al. Bulk elemental analysis measurements confirm that hypothesis: the percent oxygen composition was 5-fold higher for the contaminated site on June 30 at Pensacola Beach (4.70 ± 0.06%) than for the MWO (0.93 ± 0.01%). The low abundance of O1 and high abundance of O≥2 classes for negative ion (−) ESI and APPI may be attributed to preferential deprotonation of acids. That result is corroborated by the Ox class isoabundance-contoured DBE vs carbon number plots for PB 5.5−7.0 cm shown in Figure 3B. A DBE vs carbon number plot is a typical way to visualize continuous petroleum compositional distributions from highresolution mass spectrometric data with respect to heteroatomic class (N1, N1O1, O1, S1, etc.).28 In contrast, the distributions of Ox classes observed by positive ESI and APPI FT-ICR MS class-specific compositional images for O1−O8 classes reveal compositionally diverse Ox classes (C25−C80 and DBE = 3−30), decreasing in relative abundance from the lowest (O1) to highest (O8) oxygen number (Figure 3A and 4A). The mono-oxygenated species observed as positive ions are likely alcohols or ketones, whereas the dioxygenated species 4046

dx.doi.org/10.1021/ef500427n | Energy Fuels 2014, 28, 4043−4050

Energy & Fuels

Article

Figure 4. Macondo well oil (green) and June 30, 2010 Pensacola Beach sediment extract (5.5−7.0 cm) (red) heteroatom class distribution for (±) APPI. Oxygenated class isoabundance-contoured DBE vs carbon number plots are shown below each heteroatom class distribution for both (+) and (−) APPI.

fragmentation.31 Therefore, ketones were targeted for isolation, enrichment, and subsequent GC×GC/TOF MS analysis. Anion-Exchange Separation for Ketone Enrichment. To validate the chemical functionality (ketone, alcohol, ester, ether) of oxygen for positive ion electrospray, we performed an anion-exchange polarity-based separation of ketones and carboxylic acids for the PB 5.5−7.0 cm extract, followed by FT-ICR MS analysis. The oxygen (Ox) heteroatom class distribution (>0.5% relative abundance) for the positive ESI FT-ICR mass analysis of the anion-exchange fractions is shown in Figure S4 (Supporting Information). The first fraction ("saturates" eluted with hexane) comprises the highest weight percent of the contaminated sand (11.3 mg, 56%). Although that fraction contains mostly saturated hydrocarbons, it still contains highly abundant O1 and O2 classes (likely high molecular weight ketones and diketones that are weakly retained by the anion-exchange resin) but a lower relative abundance of higher-order oxygenated compounds. Most of the ketones (as determined by GC×GC TOF/MS) were eluted by 100% dichloromethane (DCM) in the second fraction (5.8 mg, 29%). Higher-order oxygenated compounds elute with 90:10 DCM/MeOH in the third fraction (2.8 mg, 14%) and contain even more oxygen atoms. Finally, the fourth fraction (0.8 mg, 4%), eluted with 90:10 methanol/acetic acid, contains the acids (confirmed by FT-ICR MS analysis), which are retained on the column by strong interactions and oxygen species up to O8. The efficiency of the separation is low due to the large number of mixed oxygen functionality species in the sample. For

that ketones are much more efficiently ionized by (+) ESI. A similar abundance for pyridinic-nitrogen-containing species and ketones is observed at low concentration of Pensacola Beach sediment extract by positive ion FT-ICR MS (see Figure 5, inset). Thus, we hypothesize that ketones are principally responsible for the higher abundance of oxygenated species observed in positive ESI analysis of the PB 5.5−7.0 cm extract. That hypothesis is further corroborated by results obtained with an equimolar model compound mixture spiked into the PB 5.5−7.0 and directly infused into the FT-ICR mass analyzer (data not shown). Biochemical support for the hypothesis is provided by the knowledge that aerobic biodegradation of alkanes (or alkyl chains) proceeds through terminal or subterminal oxidation.30 Terminal oxidation begins with the oxidation of a terminal methyl group to yield a primary alcohol that is subsequently converted (in multiple steps) to a fatty acid. The acids are readily identified here by (−) ESI FT-ICR MS analysis and account for the sudden onset of highly abundant, acidic species at the O2 class. Subterminal oxidation results in a secondary alcohol that is further dehydrogenated by alcohol dehydrogenase to yield the corresponding ketone. Combined with the results in the analysis of standard compounds (presented above), such a process could explain the sudden onset of abundant species in the (+) ESI FT-ICR mass spectrum at the O1 class. If correct, the analysis of the beach sand extracts (or ketone-enriched fractions) by comprehensive 2-D gas chromatography/mass spectrometry should yield abundant ketone fragment ions of m/z = 58 (C3H6O1), indicative of alkyl 2-one 4047

dx.doi.org/10.1021/ef500427n | Energy Fuels 2014, 28, 4043−4050

Energy & Fuels

Article

Figure 5. GC×GC/MS selected positive ion mass chromatograms (m/z 58) for June 30, 2010 Pensacola Beach sediment extract (5.5−7.0 cm depth) anion-exchange LC fractions 1−3.

alkanes (a commonly observed alkane fragment is m/z 57) that cannot be resolved by low-resolution (nominal mass resolution) TOF MS. However, because of extensive weathering highlighted in Figure 1, 13C n-alkanes in those two fractions are lost to evaporation and are thus not present in the beach sediment extract. It is worth noting that C22−C52 alkyl ketones are tentatively identified by (+) ESI FT-ICR MS broadband analysis (Figure 3A, O1 compositional image) of the whole beach sediment extract (a zoom inset for the low (0−5) DBE species is provided in Figure S5, Supporting Information) but ultimately confirmed by the GC×GC TOF/MS analysis by a match of the 70 eV mass spectrum to the NIST mass spectral database. Thus, future analyses of contaminated sediments, tar balls, sand patties, rock scrapings, and other field samples should yield direct characterization of highly oxidized species without chromatographic enrichment (performed here for verification purposes). Future Analytical Needs. The source (biotic or abiotic) of the oxidative products identified herein is an important step for understanding the available chemical transformation pathways for petrogenic species released into the environment. Oxidation of petrogenic species in the environment is reported to proceed through alkane functionalities (either distinct or as an alkyl side chain).33 To confirm that the oxidative products observed by (+) ESI FT-ICR MS originate from alkyl side chains (which would thus be lost by tandem MS experiments that lead to dealkylation), we performed infrared multiphoton dissociation (IRMPD) tandem MS experiments on both the parent petroleum and sediment extracts. Figure 6 displays heteroatom class distributions from isolated mass segments (blue) and associated IRMPD spectra (red) for Macondo crude (top) and Pensacola Beach sediment extract (bottom) samples. At identical laser power and irradiation period, the relative

example, an O3 species could be composed of a combination of hydroxyl, ketone, and carboxylic functionalities. GC×GC/TOF MS Analysis. To confirm the structural assignments supported by FT-ICR MS and TOF MS results, we performed GC×GC/TOF MS analysis of the unfractionated and fractionated PB 5.5−7.0 cm extract. We analyzed the first three chromatographic fractions (the fourth fraction contains carboxylic acids that are not amenable to GC analysis without derivatization). Hydrocarbon dominance (alkanes and cycloalkanes) is demonstrated by the GC×GC selected ion chromatogram for ions of m/z = 58 (the 13C isotopologue of abundant alkane fragments at m/z = 57) for the first anionexchange chromatography fraction (Figure 5, top) that is nearly identical to the unfractionated PB beach extract (data not shown). Although present, the high carbon number, alkyl ketone fragments overlap with (and are thus partially masked by) the 13C isotopologue (m/z = 58) of abundant alkane fragments whose monoisotopic mass lies at m/z = 57. In contrast, a selected ion mass chromatogram for ions of m/z 58 (Figure 5, middle) reveals an easily identifiable homologous series of alkyl ketones (n-alkane-2-ones).32 The ions with nalkane carbon numbers less than 15 are, however, most likely not native MWO n-alkanes, because they evaporate during weathering.4 The second fraction is preferentially enriched in high carbon number ketones (C19−23 most abundant) relative to the third fraction (Figure 5, bottom, in which C7−12 ions are most abundant) due to their weaker interaction with the liquid chromatographic stationary phase (a result of their higher carbon number) and higher polarity of the eluting solvent. The m/z 58 mass trace for those two fractions enabled mass spectrometric identification of ketones. Note, however, that the assignment of the n-alkane-2-ones would be ambiguous in unweathered crudes due to the 13C isotopologue of abundant 4048

dx.doi.org/10.1021/ef500427n | Energy Fuels 2014, 28, 4043−4050

Energy & Fuels

Article

hydrocarbon (loss of an oxygen from an O1 class generates a member of the hydrocarbon class) or Ox−1 class (i.e., loss of oxygen from an O2 class generates a member of the O1 class) during fragmentation. Although limited to a single contamination site, this study is part of a Gulf of Mexico-wide effort to understand the effects of the oil spill. Similar oxygen-containing species are found all along the Gulf and will be the subject of future manuscripts. Overall, the combination of FT-ICR-MS with GC×GC and fractionation of samples based on their polarity provides valuable insights into oxygenation of petroleum hydrocarbons.

■

ASSOCIATED CONTENT

S Supporting Information *

Experimental methods and figures. This material is available free of charge via the Internet at http://pubs.acs.org.

■

Figure 6. Heteroatom class distributions from positive ion quadrupole-isolated ESI FT-ICR mass spectra before (blue) and after (red) IRMPD fragmentation, for Macondo wellhead (top) and Pensacola Beach (bottom) PB 5.5−7.0 cm sediment extract samples.

AUTHOR INFORMATION

Corresponding Author

*Phone: +1 850 644 2398. Fax: +1 850 644 1366. E-mail: [email protected]. Notes

abundance for hydrocarbon (HC, far right on chart) dissociation products is much higher for the PB 5.5−7.0 cm extract than for the Macondo wellhead oil. Thus, the positive ion ESI-detected oxygen modifications are localized to alkyl side chains and not in, or immediately adjacent to, aromatic rings (that would fragment without a change in heteroatom class). Although such modifications are consistent with biotic transformation products (subterminal alkyl biooxidation), additional studies that employ biotic and abiotic control microcosms will be required to discern whether or not MSn methods can differentiate between biotic and abiotic transformation mechanisms.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The work was supported by the National Science Foundation (NSF) Division of Materials Research through DMR-06-54118, NSF Division of Chemistry through Rapid Grant CHE1049753, the State of Florida, NSF grants OCE-1044939 and OCE-1057417, and BP/The Gulf of Mexico Research Initiative to the Deep-C Consortium. The authors thank Jacqueline M. Jarvis for the elemental analysis measurements and Christopher L. Hendrickson for helpful discussions.

■

■

CONCLUSIONS Upon oil weathering, various oxygenated compounds are formed. Components of the oxygenated fraction in weathered oil have often been overlooked in oil spill studies due to the analytical challenges associated with their increased polarity (boiling point) and compositional complexity. Although it has recently been realized that they form rapidly during oil weathering,4 their comprehensive characterization on a molecular level has not yet been presented. We were able to identify tens of thousands of compounds formed upon oil weathering, containing one to eight oxygen atoms per molecule. Among the formed products, we identified ketones and carboxylic acids. Ketone transformation products can be readily identified by (+) ESI, whereas acidic transformation products are detected by (−) ESI. Chromatographic fractionation of the oxygenated species yields an overlapping class progression from O1−O2 ketones in the first fraction, O1−O5 species in the second fraction, O1−O7 species in the third, and polyfunctional ketone/carboxylic acid species (O1−O8) in the fourth fraction. The elution order is a function of oxygen content/functionality (higher oxygen content/acids elute later) and molecular weight (lower molecular weight species elute later); thus, the polyfunctionality (ketone and acid) of the higher oxygencontaining species is accessed by both (±) ESI FT-ICR MS. The formation of alcohols is also likely, but they are potentially suppressed by the more efficiently ionized ketones. Tandem MS experiments suggest that the oxygen incorporation is localized to alkyl chains and is thereby lost to yield a

REFERENCES

(1) McNutt, M. K.; Camilli, R.; Crone, T. J.; Guthrie, G. D.; Hsieh, P. A.; Ryerson, T. B.; Savas, O.; Shaffer, F. Proc. Natl. Acad. Sci. U.S.A. 2011, www.pnas.org/cgi/doi/10.1073/pnas.1112139108. (2) Graham, B.; Reilly, W. K.; Beinecke, F.; Boesch, D. F.; Garcia, T. D.; Murray, C. A. Deep Water: The Gulf Oil Disaster and the Future of Offshore Drilling; Report to the President: National Commission on the BP Deepwater Horizon Oil Spill Offshore Drilling; U.S. Government Printing Office: Washington, DC, 2011. (3) Atlas, R. M. Environ. Int. 1981, 5, 33−38. (4) Aeppli, C.; Carmichael, C. A.; Nelson, R. K.; Lemkau, K. L.; Graham, W. M.; Redmond, M. C.; Valentine, D. L.; Reddy, C. M. Environ. Sci. Technol. 2012, 46, 8799−8807. (5) Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S. Mass Spectrom. Rev. 1998, 17, 1−35. (6) Kaiser, N. K.; Savory, J. J.; McKenna, A. M.; Quinn, J. P.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2011, 83, 6907−6910. (7) Podgorski, D. C.; Corilo, Y. E.; Nyadong, L.; Lobodin, V. V.; Bythell, B. J.; Robbins, W. K.; McKenna, A. M.; Marshall, A. G.; Rodgers, R. P. Energy Fuels 2013, 27, 1268−1276. (8) Savory, J. J.; Kaiser, N. K.; McKenna, A. M.; Xian, F.; Blakney, G. T.; Rodgers, R. P.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2011, 83, 1732−1736. (9) Qian, K.; Robbins, W. K.; Hughey, C. A.; Cooper, H. J.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2001, 15, 1505−1511. (10) Mapolelo, M. M.; Stanford, L. A.; Rodgers, R. P.; Yen, A. T.; Debord, J. D.; Asomaning, S.; Marshall, A. G. Energy Fuels 2009, 23, 349−355. (11) Qian, K.; Rodgers, R. P.; Hendrickson, C. L.; Emmett, M.; Marshall, A. G. Energy Fuels 2001, 15, 492−498. 4049

dx.doi.org/10.1021/ef500427n | Energy Fuels 2014, 28, 4043−4050

Energy & Fuels

Article

(12) (a) O’Connor, P. B.; McLafferty, F. W. J. Am. Soc. Mass. Spectrom. 1995, 6, 533−535. (b) Chen, L.; Marshall, A. G. Int. J. Mass Spectrom. Ion Processes 1987, 79, 115−125. (13) Laskin, J.; Futrell, J. H. Mass Spectrom. Rev. 2003, 22, 158−181. (14) Solouki, T.; Freitas, M. A.; Alomary, A. Anal. Chem. 1999, 71, 4719−4726. (15) (a) Marshall, A. G.; Rodgers, R. P. Acc. Chem. Res. 2004, 37, 53−59. (b) Marshall, A. G.; Rodgers, R. P. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 18090−18095. (c) Rodgers, R. P.; Schaub, T. C.; Marshall, A. G. Anal. Chem. 2005, 77, 20A−27A. (16) (a) McKenna, A. M.; Marshall, A. G.; Rodgers, R. P. Energy Fuels 2013, 27, 1257−1267. (b) McKenna, A. M.; Donald, L. J.; Fitzsimmons, J. E.; Juyal, P.; Spicer, V.; Standing, K. G.; Marshall, A. G.; Rodgers, R. P. Energy Fuels 2013, 27, 1246−1256. (17) McKenna, A. M.; Nelson, R. K.; Reddy, C. M.; Savory, J. J.; Kaiser, N. K.; Fitzsimmons, J. E.; Marshall, A. G.; Rodgers, R. P. Environ. Sci. Technol. 2013, 47, 7530−7539. (18) Hughey, C. A.; Minardi, C. S.; Galasso-Roth, S. A.; B, P. G.; Mapolelo, M. M.; Rodgers, R. P.; Marshall, A. G.; Ruderman, D. L. Rapid Commun. Mass Spectrom. 2008, 22, 3968−3976. (19) (a) Rodgers, R. P.; Blumer, E. N.; Freitas, M. A.; Marshall, A. G. Anal. Chem. 1999, 71, 5171−5176. (b) Rodgers, R. P.; Blumer, E. N.; Freitas, M. A.; Marshall, A. G. Environ. Sci. Technol. 2000, 34, 1671− 1678. (20) Payne, J.; Phillips, C. Environ. Sci. Technol. 1985, 19, 569−579. (21) Carmichael, C. A.; Arey, J. S.; Graham, W. M.; Linn, L. J.; Lemkau, K. L.; Nelson, R. K.; Reddy, C. M. Environ. Res. Lett. 2012, 7, 015301. (22) Kostka, J. E.; Prakash, O.; Overholt, W. A.; Green, S. J.; Freyer, G.; Canion, A.; Delgardio, J.; Norton, N.; Hazen, T. C.; Huettel, M. Appl. Environ. Microbiol. 2011, 77, 7962−7974. (23) Liu, Z.; Liu, J.; Zhu, Q.; Wu, W. Environ. Res. Lett. 2012, 7, 035302. (24) McLafferty, F. W.; Turecek, F. Interpretation of Mass Spectra, 4th ed.; University Science Books: Mill Valley, CA, 1993. (25) Schaub, T. M.; Jennings, D. W.; Kim, S.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2007, 21, 185−194. Barrow, M. P.; Witt, M.; Headly, J. V.; Peru, K. M. Anal. Chem. 2010, 82, 3727−3735. (26) Marshall, A. G.; Wang, T.-C. L.; Ricca, T. L. J. Am. Soc. Mass. Spectrom. 1985, 107, 7893−7897. Wang, T. C. L.; Ricca, T. L.; Marshall, A. G. Anal. Chem. 1986, 58, 2935−2938. (27) Ruddy, B. M.; Henrickson, C. L.; Rodgers, R. P.; Marshall, A. G. In Proceedings of the 60th American Society for Mass Spectrometry Conference on Mass Spectrometry and Allied Topics, Vancouver, Brittish Colombia, Canada, May 20−24, 2012; American Society for Mass Spectrometry: Santa Fe, NM, 2012. (28) Hughey, C. A.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G.; Qian, K. Anal. Chem. 2001, 4676−4681. (29) Mapolelo, M. M.; Rodgers, R. P.; Blakney, G. T.; Yen, A. T.; Asomaning, S.; Marshall, A. G. Int. J. Mass Spectrom. 2011, 300, 149− 157. (30) Rojo, F. Environ. Microbiol. 2009, 11, 2477−2490. (31) Zhou, J.; Zong, Z.-M.; Chen, B.; Yang, Z. S.; Li, P.; Lu, Y.; Yue, X.-M.; Cong, X.-S.; Wei, Y.-B.; Wang, Y.-G; Fan, X.; Zhao, Y.-P.; Wei, X.-Y. Energy Sources, Part A 2013, 35, 2218−2224. (32) Strel’nikova, E. B.; Stakhina, L. D.; Petrenko, T. V. J. Anal. Chem. 2009, 64, 8−13. (33) Rontani, J.-F.; Giral, P. J.-P. Int. J. Environ. Anal. Chem. 1990, 42, 61−68. Guillano, M.; El Anba-Lurot, F.; Doumenq, P.; Mille, G.; Rontani, J. F. J. Photochem. Photobiol., A 1997, 102, 127−132.

4050

dx.doi.org/10.1021/ef500427n | Energy Fuels 2014, 28, 4043−4050